PrintPreprocessedOutput.cpp revision 44fad217d390a45a6e927140bfcbd59c5f4fbab5 
1//===--- PrintPreprocessedOutput.cpp - Implement the -E mode --------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This code simply runs the preprocessor on the input file and prints out the
11// result.  This is the traditional behavior of the -E option.
12//
13//===----------------------------------------------------------------------===//
14
15#include "clang/Frontend/Utils.h"
16#include "clang/Basic/Diagnostic.h"
17#include "clang/Basic/SourceManager.h"
18#include "clang/Frontend/PreprocessorOutputOptions.h"
19#include "clang/Lex/MacroInfo.h"
20#include "clang/Lex/PPCallbacks.h"
21#include "clang/Lex/Pragma.h"
22#include "clang/Lex/Preprocessor.h"
23#include "clang/Lex/TokenConcatenation.h"
24#include "llvm/ADT/SmallString.h"
25#include "llvm/ADT/STLExtras.h"
26#include "llvm/ADT/StringRef.h"
27#include "llvm/Config/config.h"
28#include "llvm/Support/raw_ostream.h"
29#include <cstdio>
30using namespace clang;
31
32/// PrintMacroDefinition - Print a macro definition in a form that will be
33/// properly accepted back as a definition.
34static void PrintMacroDefinition(const IdentifierInfo &II, const MacroInfo &MI,
35                                 Preprocessor &PP, llvm::raw_ostream &OS) {
36  OS << "#define " << II.getName();
37
38  if (MI.isFunctionLike()) {
39    OS << '(';
40    if (!MI.arg_empty()) {
41      MacroInfo::arg_iterator AI = MI.arg_begin(), E = MI.arg_end();
42      for (; AI+1 != E; ++AI) {
43        OS << (*AI)->getName();
44        OS << ',';
45      }
46
47      // Last argument.
48      if ((*AI)->getName() == "__VA_ARGS__")
49        OS << "...";
50      else
51        OS << (*AI)->getName();
52    }
53
54    if (MI.isGNUVarargs())
55      OS << "...";  // #define foo(x...)
56
57    OS << ')';
58  }
59
60  // GCC always emits a space, even if the macro body is empty.  However, do not
61  // want to emit two spaces if the first token has a leading space.
62  if (MI.tokens_empty() || !MI.tokens_begin()->hasLeadingSpace())
63    OS << ' ';
64
65  llvm::SmallString<128> SpellingBuffer;
66  for (MacroInfo::tokens_iterator I = MI.tokens_begin(), E = MI.tokens_end();
67       I != E; ++I) {
68    if (I->hasLeadingSpace())
69      OS << ' ';
70
71    OS << PP.getSpelling(*I, SpellingBuffer);
72  }
73}
74
75//===----------------------------------------------------------------------===//
76// Preprocessed token printer
77//===----------------------------------------------------------------------===//
78
79namespace {
80class PrintPPOutputPPCallbacks : public PPCallbacks {
81  Preprocessor &PP;
82  SourceManager &SM;
83  TokenConcatenation ConcatInfo;
84public:
85  llvm::raw_ostream &OS;
86private:
87  unsigned CurLine;
88  bool EmittedTokensOnThisLine;
89  bool EmittedMacroOnThisLine;
90  SrcMgr::CharacteristicKind FileType;
91  llvm::SmallString<512> CurFilename;
92  bool Initialized;
93  bool DisableLineMarkers;
94  bool DumpDefines;
95  bool UseLineDirective;
96public:
97  PrintPPOutputPPCallbacks(Preprocessor &pp, llvm::raw_ostream &os,
98                           bool lineMarkers, bool defines)
99     : PP(pp), SM(PP.getSourceManager()),
100       ConcatInfo(PP), OS(os), DisableLineMarkers(lineMarkers),
101       DumpDefines(defines) {
102    CurLine = 0;
103    CurFilename += "<uninit>";
104    EmittedTokensOnThisLine = false;
105    EmittedMacroOnThisLine = false;
106    FileType = SrcMgr::C_User;
107    Initialized = false;
108
109    // If we're in microsoft mode, use normal #line instead of line markers.
110    UseLineDirective = PP.getLangOptions().Microsoft;
111  }
112
113  void SetEmittedTokensOnThisLine() { EmittedTokensOnThisLine = true; }
114  bool hasEmittedTokensOnThisLine() const { return EmittedTokensOnThisLine; }
115
116  virtual void FileChanged(SourceLocation Loc, FileChangeReason Reason,
117                           SrcMgr::CharacteristicKind FileType);
118  virtual void Ident(SourceLocation Loc, const std::string &str);
119  virtual void PragmaComment(SourceLocation Loc, const IdentifierInfo *Kind,
120                             const std::string &Str);
121  virtual void PragmaMessage(SourceLocation Loc, llvm::StringRef Str);
122
123  bool HandleFirstTokOnLine(Token &Tok);
124  bool MoveToLine(SourceLocation Loc) {
125    return MoveToLine(SM.getPresumedLoc(Loc).getLine());
126  }
127  bool MoveToLine(unsigned LineNo);
128
129  bool AvoidConcat(const Token &PrevPrevTok, const Token &PrevTok,
130                   const Token &Tok) {
131    return ConcatInfo.AvoidConcat(PrevPrevTok, PrevTok, Tok);
132  }
133  void WriteLineInfo(unsigned LineNo, const char *Extra=0, unsigned ExtraLen=0);
134
135  void HandleNewlinesInToken(const char *TokStr, unsigned Len);
136
137  /// MacroDefined - This hook is called whenever a macro definition is seen.
138  void MacroDefined(const IdentifierInfo *II, const MacroInfo *MI);
139
140};
141}  // end anonymous namespace
142
143void PrintPPOutputPPCallbacks::WriteLineInfo(unsigned LineNo,
144                                             const char *Extra,
145                                             unsigned ExtraLen) {
146  if (EmittedTokensOnThisLine || EmittedMacroOnThisLine) {
147    OS << '\n';
148    EmittedTokensOnThisLine = false;
149    EmittedMacroOnThisLine = false;
150  }
151
152  // Emit #line directives or GNU line markers depending on what mode we're in.
153  if (UseLineDirective) {
154    OS << "#line" << ' ' << LineNo << ' ' << '"';
155    OS.write(&CurFilename[0], CurFilename.size());
156    OS << '"';
157  } else {
158    OS << '#' << ' ' << LineNo << ' ' << '"';
159    OS.write(&CurFilename[0], CurFilename.size());
160    OS << '"';
161
162    if (ExtraLen)
163      OS.write(Extra, ExtraLen);
164
165    if (FileType == SrcMgr::C_System)
166      OS.write(" 3", 2);
167    else if (FileType == SrcMgr::C_ExternCSystem)
168      OS.write(" 3 4", 4);
169  }
170  OS << '\n';
171}
172
173/// MoveToLine - Move the output to the source line specified by the location
174/// object.  We can do this by emitting some number of \n's, or be emitting a
175/// #line directive.  This returns false if already at the specified line, true
176/// if some newlines were emitted.
177bool PrintPPOutputPPCallbacks::MoveToLine(unsigned LineNo) {
178  // If this line is "close enough" to the original line, just print newlines,
179  // otherwise print a #line directive.
180  if (LineNo-CurLine <= 8) {
181    if (LineNo-CurLine == 1)
182      OS << '\n';
183    else if (LineNo == CurLine)
184      return false;    // Spelling line moved, but instantiation line didn't.
185    else {
186      const char *NewLines = "\n\n\n\n\n\n\n\n";
187      OS.write(NewLines, LineNo-CurLine);
188    }
189  } else if (!DisableLineMarkers) {
190    // Emit a #line or line marker.
191    WriteLineInfo(LineNo, 0, 0);
192  } else {
193    // Okay, we're in -P mode, which turns off line markers.  However, we still
194    // need to emit a newline between tokens on different lines.
195    if (EmittedTokensOnThisLine || EmittedMacroOnThisLine) {
196      OS << '\n';
197      EmittedTokensOnThisLine = false;
198      EmittedMacroOnThisLine = false;
199    }
200  }
201
202  CurLine = LineNo;
203  return true;
204}
205
206
207/// FileChanged - Whenever the preprocessor enters or exits a #include file
208/// it invokes this handler.  Update our conception of the current source
209/// position.
210void PrintPPOutputPPCallbacks::FileChanged(SourceLocation Loc,
211                                           FileChangeReason Reason,
212                                       SrcMgr::CharacteristicKind NewFileType) {
213  // Unless we are exiting a #include, make sure to skip ahead to the line the
214  // #include directive was at.
215  SourceManager &SourceMgr = SM;
216
217  PresumedLoc UserLoc = SourceMgr.getPresumedLoc(Loc);
218  unsigned NewLine = UserLoc.getLine();
219
220  if (Reason == PPCallbacks::EnterFile) {
221    SourceLocation IncludeLoc = SourceMgr.getPresumedLoc(Loc).getIncludeLoc();
222    if (IncludeLoc.isValid())
223      MoveToLine(IncludeLoc);
224  } else if (Reason == PPCallbacks::SystemHeaderPragma) {
225    MoveToLine(NewLine);
226
227    // TODO GCC emits the # directive for this directive on the line AFTER the
228    // directive and emits a bunch of spaces that aren't needed.  Emulate this
229    // strange behavior.
230  }
231
232  CurLine = NewLine;
233
234  if (DisableLineMarkers) return;
235
236  CurFilename.clear();
237  CurFilename += UserLoc.getFilename();
238  Lexer::Stringify(CurFilename);
239  FileType = NewFileType;
240
241  if (!Initialized) {
242    WriteLineInfo(CurLine);
243    Initialized = true;
244  }
245
246  switch (Reason) {
247  case PPCallbacks::EnterFile:
248    WriteLineInfo(CurLine, " 1", 2);
249    break;
250  case PPCallbacks::ExitFile:
251    WriteLineInfo(CurLine, " 2", 2);
252    break;
253  case PPCallbacks::SystemHeaderPragma:
254  case PPCallbacks::RenameFile:
255    WriteLineInfo(CurLine);
256    break;
257  }
258}
259
260/// Ident - Handle #ident directives when read by the preprocessor.
261///
262void PrintPPOutputPPCallbacks::Ident(SourceLocation Loc, const std::string &S) {
263  MoveToLine(Loc);
264
265  OS.write("#ident ", strlen("#ident "));
266  OS.write(&S[0], S.size());
267  EmittedTokensOnThisLine = true;
268}
269
270/// MacroDefined - This hook is called whenever a macro definition is seen.
271void PrintPPOutputPPCallbacks::MacroDefined(const IdentifierInfo *II,
272                                            const MacroInfo *MI) {
273  // Only print out macro definitions in -dD mode.
274  if (!DumpDefines ||
275      // Ignore __FILE__ etc.
276      MI->isBuiltinMacro()) return;
277
278  MoveToLine(MI->getDefinitionLoc());
279  PrintMacroDefinition(*II, *MI, PP, OS);
280  EmittedMacroOnThisLine = true;
281}
282
283
284void PrintPPOutputPPCallbacks::PragmaComment(SourceLocation Loc,
285                                             const IdentifierInfo *Kind,
286                                             const std::string &Str) {
287  MoveToLine(Loc);
288  OS << "#pragma comment(" << Kind->getName();
289
290  if (!Str.empty()) {
291    OS << ", \"";
292
293    for (unsigned i = 0, e = Str.size(); i != e; ++i) {
294      unsigned char Char = Str[i];
295      if (isprint(Char) && Char != '\\' && Char != '"')
296        OS << (char)Char;
297      else  // Output anything hard as an octal escape.
298        OS << '\\'
299           << (char)('0'+ ((Char >> 6) & 7))
300           << (char)('0'+ ((Char >> 3) & 7))
301           << (char)('0'+ ((Char >> 0) & 7));
302    }
303    OS << '"';
304  }
305
306  OS << ')';
307  EmittedTokensOnThisLine = true;
308}
309
310void PrintPPOutputPPCallbacks::PragmaMessage(SourceLocation Loc,
311                                             llvm::StringRef Str) {
312  MoveToLine(Loc);
313  OS << "#pragma message(";
314
315  OS << '"';
316
317  for (unsigned i = 0, e = Str.size(); i != e; ++i) {
318    unsigned char Char = Str[i];
319    if (isprint(Char) && Char != '\\' && Char != '"')
320      OS << (char)Char;
321    else  // Output anything hard as an octal escape.
322      OS << '\\'
323         << (char)('0'+ ((Char >> 6) & 7))
324         << (char)('0'+ ((Char >> 3) & 7))
325         << (char)('0'+ ((Char >> 0) & 7));
326  }
327  OS << '"';
328
329  OS << ')';
330  EmittedTokensOnThisLine = true;
331}
332
333
334/// HandleFirstTokOnLine - When emitting a preprocessed file in -E mode, this
335/// is called for the first token on each new line.  If this really is the start
336/// of a new logical line, handle it and return true, otherwise return false.
337/// This may not be the start of a logical line because the "start of line"
338/// marker is set for spelling lines, not instantiation ones.
339bool PrintPPOutputPPCallbacks::HandleFirstTokOnLine(Token &Tok) {
340  // Figure out what line we went to and insert the appropriate number of
341  // newline characters.
342  if (!MoveToLine(Tok.getLocation()))
343    return false;
344
345  // Print out space characters so that the first token on a line is
346  // indented for easy reading.
347  unsigned ColNo = SM.getInstantiationColumnNumber(Tok.getLocation());
348
349  // This hack prevents stuff like:
350  // #define HASH #
351  // HASH define foo bar
352  // From having the # character end up at column 1, which makes it so it
353  // is not handled as a #define next time through the preprocessor if in
354  // -fpreprocessed mode.
355  if (ColNo <= 1 && Tok.is(tok::hash))
356    OS << ' ';
357
358  // Otherwise, indent the appropriate number of spaces.
359  for (; ColNo > 1; --ColNo)
360    OS << ' ';
361
362  return true;
363}
364
365void PrintPPOutputPPCallbacks::HandleNewlinesInToken(const char *TokStr,
366                                                     unsigned Len) {
367  unsigned NumNewlines = 0;
368  for (; Len; --Len, ++TokStr) {
369    if (*TokStr != '\n' &&
370        *TokStr != '\r')
371      continue;
372
373    ++NumNewlines;
374
375    // If we have \n\r or \r\n, skip both and count as one line.
376    if (Len != 1 &&
377        (TokStr[1] == '\n' || TokStr[1] == '\r') &&
378        TokStr[0] != TokStr[1])
379      ++TokStr, --Len;
380  }
381
382  if (NumNewlines == 0) return;
383
384  CurLine += NumNewlines;
385}
386
387
388namespace {
389struct UnknownPragmaHandler : public PragmaHandler {
390  const char *Prefix;
391  PrintPPOutputPPCallbacks *Callbacks;
392
393  UnknownPragmaHandler(const char *prefix, PrintPPOutputPPCallbacks *callbacks)
394    : PragmaHandler(0), Prefix(prefix), Callbacks(callbacks) {}
395  virtual void HandlePragma(Preprocessor &PP, Token &PragmaTok) {
396    // Figure out what line we went to and insert the appropriate number of
397    // newline characters.
398    Callbacks->MoveToLine(PragmaTok.getLocation());
399    Callbacks->OS.write(Prefix, strlen(Prefix));
400
401    // Read and print all of the pragma tokens.
402    while (PragmaTok.isNot(tok::eom)) {
403      if (PragmaTok.hasLeadingSpace())
404        Callbacks->OS << ' ';
405      std::string TokSpell = PP.getSpelling(PragmaTok);
406      Callbacks->OS.write(&TokSpell[0], TokSpell.size());
407      PP.LexUnexpandedToken(PragmaTok);
408    }
409    Callbacks->OS << '\n';
410  }
411};
412} // end anonymous namespace
413
414
415static void PrintPreprocessedTokens(Preprocessor &PP, Token &Tok,
416                                    PrintPPOutputPPCallbacks *Callbacks,
417                                    llvm::raw_ostream &OS) {
418  char Buffer[256];
419  Token PrevPrevTok, PrevTok;
420  PrevPrevTok.startToken();
421  PrevTok.startToken();
422  while (1) {
423
424    // If this token is at the start of a line, emit newlines if needed.
425    if (Tok.isAtStartOfLine() && Callbacks->HandleFirstTokOnLine(Tok)) {
426      // done.
427    } else if (Tok.hasLeadingSpace() ||
428               // If we haven't emitted a token on this line yet, PrevTok isn't
429               // useful to look at and no concatenation could happen anyway.
430               (Callbacks->hasEmittedTokensOnThisLine() &&
431                // Don't print "-" next to "-", it would form "--".
432                Callbacks->AvoidConcat(PrevPrevTok, PrevTok, Tok))) {
433      OS << ' ';
434    }
435
436    if (IdentifierInfo *II = Tok.getIdentifierInfo()) {
437      OS << II->getName();
438    } else if (Tok.isLiteral() && !Tok.needsCleaning() &&
439               Tok.getLiteralData()) {
440      OS.write(Tok.getLiteralData(), Tok.getLength());
441    } else if (Tok.getLength() < 256) {
442      const char *TokPtr = Buffer;
443      unsigned Len = PP.getSpelling(Tok, TokPtr);
444      OS.write(TokPtr, Len);
445
446      // Tokens that can contain embedded newlines need to adjust our current
447      // line number.
448      if (Tok.getKind() == tok::comment)
449        Callbacks->HandleNewlinesInToken(TokPtr, Len);
450    } else {
451      std::string S = PP.getSpelling(Tok);
452      OS.write(&S[0], S.size());
453
454      // Tokens that can contain embedded newlines need to adjust our current
455      // line number.
456      if (Tok.getKind() == tok::comment)
457        Callbacks->HandleNewlinesInToken(&S[0], S.size());
458    }
459    Callbacks->SetEmittedTokensOnThisLine();
460
461    if (Tok.is(tok::eof)) break;
462
463    PrevPrevTok = PrevTok;
464    PrevTok = Tok;
465    PP.Lex(Tok);
466  }
467}
468
469typedef std::pair<IdentifierInfo*, MacroInfo*> id_macro_pair;
470static int MacroIDCompare(const void* a, const void* b) {
471  const id_macro_pair *LHS = static_cast<const id_macro_pair*>(a);
472  const id_macro_pair *RHS = static_cast<const id_macro_pair*>(b);
473  return LHS->first->getName().compare(RHS->first->getName());
474}
475
476static void DoPrintMacros(Preprocessor &PP, llvm::raw_ostream *OS) {
477  // Ignore unknown pragmas.
478  PP.AddPragmaHandler(0, new EmptyPragmaHandler());
479
480  // -dM mode just scans and ignores all tokens in the files, then dumps out
481  // the macro table at the end.
482  PP.EnterMainSourceFile();
483
484  Token Tok;
485  do PP.Lex(Tok);
486  while (Tok.isNot(tok::eof));
487
488  llvm::SmallVector<id_macro_pair, 128>
489    MacrosByID(PP.macro_begin(), PP.macro_end());
490  llvm::array_pod_sort(MacrosByID.begin(), MacrosByID.end(), MacroIDCompare);
491
492  for (unsigned i = 0, e = MacrosByID.size(); i != e; ++i) {
493    MacroInfo &MI = *MacrosByID[i].second;
494    // Ignore computed macros like __LINE__ and friends.
495    if (MI.isBuiltinMacro()) continue;
496
497    PrintMacroDefinition(*MacrosByID[i].first, MI, PP, *OS);
498    *OS << '\n';
499  }
500}
501
502/// DoPrintPreprocessedInput - This implements -E mode.
503///
504void clang::DoPrintPreprocessedInput(Preprocessor &PP, llvm::raw_ostream *OS,
505                                     const PreprocessorOutputOptions &Opts) {
506  // Show macros with no output is handled specially.
507  if (!Opts.ShowCPP) {
508    assert(Opts.ShowMacros && "Not yet implemented!");
509    DoPrintMacros(PP, OS);
510    return;
511  }
512
513  // Inform the preprocessor whether we want it to retain comments or not, due
514  // to -C or -CC.
515  PP.SetCommentRetentionState(Opts.ShowComments, Opts.ShowMacroComments);
516
517  PrintPPOutputPPCallbacks *Callbacks =
518      new PrintPPOutputPPCallbacks(PP, *OS, !Opts.ShowLineMarkers,
519                                   Opts.ShowMacros);
520  PP.AddPragmaHandler(0, new UnknownPragmaHandler("#pragma", Callbacks));
521  PP.AddPragmaHandler("GCC", new UnknownPragmaHandler("#pragma GCC",
522                                                      Callbacks));
523
524  PP.addPPCallbacks(Callbacks);
525
526  // After we have configured the preprocessor, enter the main file.
527  PP.EnterMainSourceFile();
528
529  // Consume all of the tokens that come from the predefines buffer.  Those
530  // should not be emitted into the output and are guaranteed to be at the
531  // start.
532  const SourceManager &SourceMgr = PP.getSourceManager();
533  Token Tok;
534  do PP.Lex(Tok);
535  while (Tok.isNot(tok::eof) && Tok.getLocation().isFileID() &&
536         !strcmp(SourceMgr.getPresumedLoc(Tok.getLocation()).getFilename(),
537                 "<built-in>"));
538
539  // Read all the preprocessed tokens, printing them out to the stream.
540  PrintPreprocessedTokens(PP, Tok, Callbacks, *OS);
541  *OS << '\n';
542}
543
544